Collaborative Computing: Networking, Applications and Worksharing. 13th International Conference, CollaborateCom 2017, Edinburgh, UK, December 11–13, 2017, Proceedings

Research Article

An Electricity Power Collection Data Oriented Missing Data Imputation Solution

Download
58 downloads
  • @INPROCEEDINGS{10.1007/978-3-030-00916-8_23,
        author={Jiangqi Chen and Han Li and Ting Zhao and He Liu},
        title={An Electricity Power Collection Data Oriented Missing Data Imputation Solution},
        proceedings={Collaborative Computing: Networking, Applications and Worksharing. 13th International Conference, CollaborateCom 2017, Edinburgh, UK, December 11--13, 2017, Proceedings},
        proceedings_a={COLLABORATECOM},
        year={2018},
        month={10},
        keywords={Missing data imputation Electricity power collection data Big data KNN Data quality},
        doi={10.1007/978-3-030-00916-8_23}
    }
    
  • Jiangqi Chen
    Han Li
    Ting Zhao
    He Liu
    Year: 2018
    An Electricity Power Collection Data Oriented Missing Data Imputation Solution
    COLLABORATECOM
    Springer
    DOI: 10.1007/978-3-030-00916-8_23
Jiangqi Chen1,*, Han Li,*, Ting Zhao1,*, He Liu1,*
  • 1: Global Energy Interconnection Research Institute
*Contact email: chenjiangqi@geiri.sgcc.com.cn, lihan@ncut.edu.cn, zhaoting@geiri.sgcc.com.cn, liuhe@geiri.sgcc.com.cn

Abstract

In Smart Grid, the incompleteness of electricity power collection data is gradually prominent. Thus, this paper presents an electricity power collection big data oriented missing data imputation solution, which comprises a big data processing framework and a missing data imputation method for power consumption data. Based on big data techniques, the given framework supports the large-scale electricity power collection data acquisition, storage and processing. To get a better result, the proposed method takes advantage of the correlation between the loss rate of power and the user power consumption. The feasibility and the effectiveness of the proposed method is evaluated. Experimental results show the proposed method is able to convert incomplete data set into complete data set, and has good imputation stability. Compared with the KNN algorithm, the proposed method has a lower imputation error, and is a positive attempt to combine domain-specific algorithms with traditional algorithms.